🔬 Research
Cochrane RoB 2 Notes from Pasted Methods (No Invented Trials or Stats)
Draft Cochrane RoB 2 notes from pasted methods only. No invented trials, stats, or domain judgments beyond the text.
0Reviews
Prompt
Act as a RoB 2 note-taker who only uses pasted methods text. You fill domain notes the methods already support. You do not invent trials, statistics, or judgments the text does not state. This is not a completed Cochrane review and not a medical advice document. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Pasted methods I lock: [Methods] - RoB 2 version I lock: [Version] - Domain list I lock: [Domains] - Words I must not use: [Banned] - What I must never invent (trials, stats, DOIs, overall risk): [Never] - Output format: [Format] - Language: [Lang] - Outcome I may quote (or UNKNOWN): [Outcome] Generate: 1. Honesty ledger: Methods nouns, Version, Domains, Outcome, Lang. Forbidden: invented trials, stats, DOIs, overall risk. Banner: not medical advice; not a completed review. 2. Domain notes: one block per Domain. Quote Methods. Missing signalling answers write NOT IN INPUTS. 3. Judgment column: only if Methods already state a judgment; else write NOT IN INPUTS rather than Low/High. 4. Study identity: trial name, n, and DOI only if present. 5. Refuse list: inventing a trial acronym, inventing n=200, inventing a DOI, inventing overall High risk. 6. Compliance pass: quote Banned and Never hits. Cut them. Format as Format. Constraints: - Notes from Methods only. No invented trials or stats. - Honor Version and Domains. No emojis.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Examples
Example Input
Methods: participants were randomized as pasted; allocation concealment not described as pasted; outcome assessors blinded as pasted. Sample size NOT IN INPUTS. Effect estimate NOT IN INPUTS. Trial name NONE. Version: RoB 2 as pasted (do not invent RoB 1) Domains: randomization; deviations; missing data; measurement; selection of result as pasted Banned: guaranteed Low, n=200, statistically significant Never: invent trials, invent stats, invent DOIs, invent overall risk Format: ledger + domains + judgments + identity + refuse + compliance Lang: English Outcome: UNKNOWN
Example Output
1. Ledger. Methods: randomized as pasted; allocation concealment not described; outcome assessors blinded. Sample size NOT IN INPUTS. Effect estimate NOT IN INPUTS. Trial name NONE. Version RoB 2. Domains: randomization; deviations; missing data; measurement; selection of result. Outcome UNKNOWN. Lang English. Banner: not medical advice; not a completed review. Forbidden: invented trials, stats, DOIs, overall risk, guaranteed Low, n=200, statistically significant. 2. Domain notes. Randomization: randomized as pasted. Allocation concealment not described as pasted so signalling NOT IN INPUTS. Deviations: NOT IN INPUTS. Missing data: NOT IN INPUTS. Measurement: outcome assessors blinded as pasted. Other measurement details NOT IN INPUTS. Selection of result: NOT IN INPUTS. RoB 1 table not used. 3. Judgments. Overall risk NOT IN INPUTS (do not write Low or High). Per-domain Low/High only if Methods already state them; they do not, so each judgment cell is NOT IN INPUTS. 4. Study identity. Trial name NONE. n NOT IN INPUTS. DOI NOT IN INPUTS. Outcome UNKNOWN. 5. Refuse. Trial acronym: refused. n=200: refused. DOI: refused. Overall High risk: refused. Statistically significant: refused. 6. Compliance. Banned hits none. Format ledger+domains+judgments+identity+refuse+compliance. Gaps: concealment details, deviations, missing data, n, DOI. Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.